Most network operators and enterprises believe transformation means much more than small changes in traditional...
network services' features or prices. Changes need to be thorough or dramatic.
The internet has been the most transformational thing about communications since its inception, and it ushered in the over-the-top (OTT) business model and the experience hosting technology model. But, for network operators, transformation now means deploying carrier cloud, and the question is whether it is driving future transformation or being driven by outside forces.
If operators deploy carrier cloud to respond to a specific opportunity or competitive threat, then carrier cloud deployment is being driven by the future, not driving it. Some may be using it to proactively drive transformation.
According to some operators, declining profit per bit on connection services is driving the most interest in transformation. Five years ago, operators predicted their revenue and cost curves would probably cross in 2018, with cost exceeding revenue for the first time. While measures to cut operations costs have extended that date for most providers, crossover still seems inevitable.
So, is the response to the crossover problem driving carrier cloud deployment? And if so, how could spending more money on new infrastructure help?
The promise of NFV to lower costs
The short answer is network functions virtualization. NFV's original goal was to transfer network features from physical devices to run on virtual, hosted components. That would not only lower costs by eliminating the proprietary, expensive network components, but it would justify the deployment of carrier cloud to host virtual components. Under the NFV model, carrier cloud would drive the future.
The problem is the NFV model hasn't realized its goal. The primary focus of NFV has been single-tenant virtual customer premises equipment, and the great majority of vCPE applications use on-premises devices, not cloud resources for hosting. Furthermore, vCPE appeals primarily to Carrier Ethernet business customers, so most network elements aren't affected at all.
Other suggested missions for NFV have been primarily in mobile infrastructure built around the IP Multimedia Subsystem (IMS) and Evolved Packet Core (EPC). But IMS and EPC elements are static, multi-tenant features, and they really don't benefit from most NFV features. That doesn't mean mobile infrastructure transformation won't require carrier cloud, however -- only that it won't require NFV.
Mobile's influence on carrier cloud
The real problem is most operators are linking mobile infrastructure transformation to 5G deployment, which is unlikely to develop fully until 2020. The question is whether market events at that time will be driving carrier cloud, rather than the other way around. The primary near-term opportunities to incorporate hosted features into services still involve mobile services, but they relate more to content consumption than to service evolution.
Streaming video has exploded in many markets, driven in large part by increased interest in streaming content to smartphones and tablets. Early 5G deployments are even being linked to millimeter-wave 5G hybrids, with fiber-to-the-node being used for wireline replacement. The 5G hybrids can deliver video only in streaming form, and a shift to streaming would have a profound impact on mobile infrastructure.
Streaming video requires a major change in the content delivery networks in use today, not only to ensure video can be streamed efficiently to mobile and fixed users, but to make sure ads can be inserted correctly. The problem is exacerbated for livestreaming, because the timing of individual programs has to be constrained within the assigned slot to prevent losing the beginning of other programs.
So far, operators aren't leading a charge toward carrier cloud based on mobile streaming requirements. Certainly, they will respond to the changes as they develop, but carrier cloud deployment is lagging mobile streaming changes, not driving them. If mobile video is the only credible near-term opportunity for hosted service features, then carrier cloud is a defensive move, not a proactive one.
Carrier cloud's 'first cost' drawbacks
Operators don't doubt the need for mobile video changes, so why not take the opportunity to invest in hosting resources -- particularly at the edge, where operators enjoy an advantage in terms of available real estate? It goes back to two factors: one financial and one psychological.
If you're in profit-per-bit trouble, you are likely focusing on first cost, which is the period of deployment in which outlays for Capex and Opex exceed benefits. As the benefits develop, the net cost shifts to a net benefit. But during that first-cost period, carrier cloud would have a negative impact on operators' financials and, thus, on their stock prices. The first-cost problem means operators want to try to synchronize their early benefits with their early costs. That approach doesn't get out in front and drive the market.
The psychological problem is network operators tend to be driven by competitive risk, not opportunity. Clearly, the OTT players like Amazon, Google and Netflix dominate the new internet services market. But operators have been slow to see that market as their own and, as a result, slow to accept OTTs as competitors.
The conclusion is clear. Carrier cloud isn't driving the future; it is captive to market conditions and events. Only IoT could save it, and IoT is another opportunity still to be realized.
IoT will require a massive investment in infrastructure, considerable experience in public policy and privacy matters, and a level of event processing that favors those with available space at the network edge for server deployment. Network operators fit all of these requirements. But so far, they're infatuated by the notion that IoT will bring billions of new sensor-cell subscribers into the fold. Only a realistic service-centric view on IoT can give carriers the lead role in their own cloud deployments.